Stability AI AI Vendor Risk Profile
Open-source AI company that created Stable Diffusion and related generative models. Develops proprietary and open-weight image, video, audio, and language models while also leveraging external model components and research.
Risk overview
Risk score: 36/100
Risk tier: Moderate
Safety rating: 64/100 (higher is better)
Lower risk scores indicate lower assessed risk.
Last verified: Apr 19, 2026 Aging
Risk dimensions
| Dimension | Risk score |
|---|---|
| Data Handling | 17/100 |
| IP Exposure | 28/100 |
| Jurisdiction | 13/100 |
| Security | 73/100 |
| Regulatory Compliance | 70/100 |
| Transparency | 40/100 |
| Business Stability | 57/100 |
| Dependency Chain | 36/100 |
| Agent Governance | Not assessed |
Analyst summary
Rating: Avoid
Stability AI's Stable Diffusion is widely deployed open-weights imagery, but the company has faced serious financial distress, CEO departure, active Getty Images litigation over training data, and offers no IP indemnification. The open weights remain valuable; the company itself is a much harder recommendation.
Bottom line: Treat the open-weights models as a community resource; do not treat the company as a viable enterprise vendor today.
Strengths
- Open-weights Stable Diffusion family allows self-hosted, air-gapped deployment
- Historical role in democratizing open image generation means broad ecosystem tooling
- Open-source community continues to evolve Stable Diffusion variants independent of the company
Concerns
- Getty Images UK trial over training on 12 million Getty photos without license
- CEO Emad Mostaque resigned amid financial difficulties and reported unpaid compute bills
- Company has had multiple leadership changes and strategic pivots
- No SOC 2 Type II, limited enterprise governance disclosures
- Standard terms explicitly disclaim IP infringement liability
- LAION-5B training data has limited provenance controls; CSAM concerns documented
Best for
- Research and experimentation with Stable Diffusion open weights
- Self-hosted deployments where the company's commercial viability is not a dependency
Avoid if
- You are considering Stability AI's hosted service for commercial use
- You need a financially stable vendor for multi-year planning
- You generate commercial creative content requiring IP indemnification
- Your procurement requires demonstrated enterprise compliance posture
Citations
- Data Handling — Trains On User Data
Stability AI may use prompts and generations from its web services to improve its models, unless you use an enterprise API with a no-training agreement.
- Governance — Financial Stability
Stability AI CEO Emad Mostaque resigned amid financial difficulties and reported struggles to meet compute bills.
- Ip Profiles — Indemnification Offered
Stability AI's standard terms disclaim liability for IP infringement claims related to outputs.
- Ip Profiles — Known Ip Lawsuits
Getty Images v. Stability AI: Getty alleges Stability's Stable Diffusion was trained on 12 million Getty photos without license.
- Ip Profiles — Training Data Provenance
Stable Diffusion was trained on LAION-5B, a dataset of image-text pairs scraped from the public internet with limited provenance controls.
- Jurisdiction Profiles — Incorporation Country
Stability AI Ltd. is a UK company headquartered in London, with US subsidiary Stability AI Inc.
- Security Compliance — Gdpr Compliant
Stability AI's privacy policy references GDPR rights but offers limited enterprise DPA support.
- Security Compliance — Soc2 Type2
Stability AI has not publicly disclosed SOC 2 Type II certification as of April 2026.